Robust Short-Term Scheduling of Multiproduct Batch Plants under Demand Uncertainty

Abstract

In this paper we are developing a strategy to quantify scheduling robustness in the face of uncertainty, to increase scheduling flexibility, and to improve system performance when unexpected events occur during scheduling execution. Several robustness metrics that are commonly used in optimization literature are adopted, and new metrics are proposed that take into account the scheduling characteristics under demand uncertainty. The deterministic model for short-term scheduling proposed by Ierapetritou and Floudas (Ind. Eng. Chem. Res. 1998, 37, 4341) that is based on a continuous-time representation is used throughout this work. To improve the schedule performance, uncertainty was considered at the scheduling stage through the multiperiod programming model. It is found that the schedule from the multiperiod formulation has a much higher robustness compared to the one determined based on the most expected demand values. Moreover, this schedule is feasible over the entire expected range of uncertainty provided the vertexes of the uncertainty range are considered as scenarios, whereas the schedule corresponding to the nominal values is most commonly infeasible over some part of the uncertain range. Further flexibility analysis studies confirmed the increased feasibility of the multiperiod schedule and showed that this schedule had a much higher capacity in terms of the total production of all products as compared to the single-period schedule. Further studies are performed in order to examine the extent of correlation of improved schedule robustness (with respect to demand uncertainty) and performance under rush order arrival. Finally, to improve the schedule ability to meet rush orders at the time of order arrival, a new methodology has been proposed that generates the deterministic schedule using an iterative procedure considering the rush orders that may arrive at different times within the time horizon.